new-words

annotate new-words.py @ 41:4629e08b0d87

find changesets by author, revision, files, or words in the commit message
minifix: skip numbers as words
author Igor Chubin <igor@chub.in>
date Sun Jan 23 18:18:19 2011 +0100 (2011-01-23)
parents c3a50c0d2400
children 3ec83a7cc544
rev   line source
igor@37 1 #!/usr/bin/env python
igor@38 2 # -*- coding: utf-8 -*-
igor@37 3
igor@40 4 from __future__ import with_statement
igor@38 5 import codecs
igor@38 6 import logging
igor@38 7 import os
igor@37 8 import optparse
igor@38 9 import re
igor@38 10 import subprocess
igor@38 11 import sys
igor@38 12 import Stemmer
igor@38 13
igor@38 14 config = {
igor@38 15 'config_directory': os.environ['HOME'] + '/.new-words',
igor@38 16 'language': 'en',
igor@38 17 }
igor@38 18
igor@38 19 logging.basicConfig(filename='/tmp/new-words-py.log', level=logging.DEBUG)
igor@38 20
igor@38 21 class Normalizator:
igor@38 22 def __init__(self, language, linked_words={}):
igor@38 23 stemmer_algorithm = {
igor@38 24 'de' : 'german',
igor@38 25 'en' : 'english',
igor@38 26 'ru' : 'russian',
igor@38 27 'uk' : 'ukrainian',
igor@38 28 }
igor@38 29 self.stemmer = Stemmer.Stemmer(stemmer_algorithm[language])
igor@38 30 self.linked_words = linked_words
igor@38 31
igor@38 32 def normalize(self, word):
igor@38 33 word_chain = []
igor@38 34 while word in self.linked_words and not word in word_chain:
igor@38 35 word_chain.append(word)
igor@38 36 word = self.linked_words[word]
igor@38 37 return self.stemmer.stemWord(word.lower())
igor@37 38
igor@37 39 parser = optparse.OptionParser()
igor@37 40
igor@37 41 parser.add_option(
igor@37 42 "-a", "--no-marks",
igor@37 43 help="don't add marks (and don't save marks added by user)",
igor@37 44 action="store_true",
igor@37 45 dest="no_marks")
igor@37 46
igor@37 47 parser.add_option(
igor@37 48 "-c", "--compressed",
igor@37 49 help="show compressed wordlist: one word per group",
igor@37 50 action="store_true",
igor@37 51 dest="compressed")
igor@37 52
igor@37 53 parser.add_option(
igor@37 54 "-k", "--known-words",
igor@37 55 help="put higher words that are similar to the known words (only for English)",
igor@37 56 action="store_true",
igor@37 57 dest="compressed")
igor@37 58
igor@37 59 parser.add_option(
igor@37 60 "-l", "--language",
igor@37 61 help="specify language of text",
igor@37 62 action="store",
igor@37 63 dest="language")
igor@37 64
igor@37 65 parser.add_option(
igor@38 66 "-f", "--function",
igor@38 67 help="filter through subsystem [INTERNAL]",
igor@38 68 action="store",
igor@38 69 dest="function")
igor@38 70
igor@38 71 parser.add_option(
igor@37 72 "-m", "--merge-tag",
igor@37 73 help="merge words tagged with specified tag into the main vocabulary",
igor@37 74 action="store",
igor@37 75 dest="merge_tag")
igor@37 76
igor@37 77 parser.add_option(
igor@37 78 "-M", "--merge-tagged",
igor@37 79 help="merge words tagged with ANY tag into the main vocabulary",
igor@37 80 action="store_true",
igor@37 81 dest="merge_tagged")
igor@37 82
igor@37 83 parser.add_option(
igor@37 84 "-n", "--non-interactive",
igor@37 85 help="non-interactive mode (don't run vi)",
igor@37 86 action="store_true",
igor@37 87 dest="non_interactive")
igor@37 88
igor@37 89 parser.add_option(
igor@37 90 "-N", "--no-filter",
igor@37 91 help="switch off known words filtering",
igor@37 92 action="store_true",
igor@37 93 dest="no_filter")
igor@37 94
igor@37 95 parser.add_option(
igor@37 96 "-p", "--pages",
igor@37 97 help="work with specified pages only (pages = start-stop/total )",
igor@37 98 action="store",
igor@37 99 dest="pages")
igor@37 100
igor@37 101 parser.add_option(
igor@37 102 "-r", "--remove-tag",
igor@37 103 help="remove subvocabulary of specified tag",
igor@37 104 action="store",
igor@37 105 dest="remove_tag")
igor@37 106
igor@37 107 parser.add_option(
igor@37 108 "-s", "--text-stats",
igor@37 109 help="show the text statistics (percentage of known words and so on) and exit",
igor@37 110 action="store_true",
igor@37 111 dest="text_stats")
igor@37 112
igor@37 113 parser.add_option(
igor@37 114 "-S", "--voc-stats",
igor@37 115 help="show your vocabulary statistics (number of words and word groups)",
igor@37 116 action="store_true",
igor@37 117 dest="voc_stats")
igor@37 118
igor@37 119 parser.add_option(
igor@37 120 "-t", "--tag",
igor@37 121 help="tag known words with tag",
igor@37 122 action="store",
igor@37 123 dest="tag")
igor@37 124
igor@37 125 parser.add_option(
igor@37 126 "-T", "--show-tags",
igor@37 127 help="tag known words with tag",
igor@37 128 action="store_true",
igor@37 129 dest="show_tags")
igor@37 130
igor@37 131 parser.add_option(
igor@37 132 "-2", "--two-words",
igor@37 133 help="find 2 words' sequences",
igor@37 134 action="store_true",
igor@37 135 dest="two_words")
igor@37 136
igor@37 137 parser.add_option(
igor@37 138 "-3", "--three-words",
igor@37 139 help="find 3 words' sequences",
igor@37 140 action="store_true",
igor@37 141 dest="three_words")
igor@37 142
igor@38 143 def readlines_from_file(filename):
igor@38 144 res = []
igor@38 145 with codecs.open(filename, "r", "utf-8") as f:
igor@38 146 for line in f.readlines():
igor@38 147 res += [line]
igor@38 148 return res
igor@38 149
igor@38 150 def readlines_from_stdin():
igor@38 151 return codecs.getreader("utf-8")(sys.stdin).readlines()
igor@38 152
igor@38 153 def words_from_line(line):
igor@38 154 line = line.rstrip('\n')
igor@38 155 #return re.split('(?:\s|[*\r,.:#@()+=<>$;"?!|\[\]^%&~{}«»–])+', line)
igor@38 156 #return re.split('[^a-zA-ZäöëüßÄËÖÜß]+', line)
igor@38 157 return re.compile("(?!')(?:\W)+", flags=re.UNICODE).split(line)
igor@38 158
igor@38 159 def get_words(lines):
igor@38 160 """
igor@38 161 Returns hash of words in a file
igor@38 162 word => number
igor@38 163 """
igor@38 164 result = {}
igor@38 165 for line in lines:
igor@38 166 words = words_from_line(line)
igor@38 167 for word in words:
igor@41 168 if re.match('[0-9]*$', word):
igor@41 169 continue
igor@38 170 result.setdefault(word, 0)
igor@38 171 result[word] += 1
igor@38 172 return result
igor@38 173
igor@38 174 def load_vocabulary():
igor@38 175 return get_words(readlines_from_file("%s/%s.txt"%(config['config_directory'], config['language'])))
igor@38 176
igor@38 177 def notes_filenames():
igor@38 178 return ["%s/notes-%s.txt"%(config['config_directory'], config['language'])]
igor@38 179
igor@38 180 def load_notes(files):
igor@38 181 notes = {}
igor@38 182 for filename in files:
igor@39 183 with codecs.open(filename, "r", "utf-8") as f:
igor@38 184 for line in f.readlines():
igor@38 185 (word, note) = re.split('\s+', line.rstrip('\n'), maxsplit=1)
igor@38 186 notes.setdefault(word, {})
igor@38 187 notes[word][filename] = note
igor@38 188 return notes
igor@38 189
igor@39 190 def add_notes(lines, notes):
igor@39 191 notes_filename = notes_filenames()[0]
igor@39 192 result = []
igor@39 193 for line in lines:
igor@39 194 if line.startswith('#'):
igor@39 195 result += [line]
igor@39 196 else:
igor@39 197 match_object = re.search('^\s*\S+\s*(\S+)', line)
igor@39 198 if match_object:
igor@39 199 word = match_object.group(1)
igor@39 200 if word in notes:
igor@39 201 logging.debug(word)
igor@39 202 logging.debug(line)
igor@39 203 if notes_filename in notes[word]:
igor@39 204 line = line.rstrip('\n')
igor@39 205 line = "%-30s %s\n" % (line, notes[word][notes_filename])
igor@39 206 logging.debug(line)
igor@39 207 result += [line]
igor@39 208 else:
igor@39 209 result += [line]
igor@39 210 else:
igor@39 211 result += [line]
igor@39 212 return result
igor@39 213
igor@39 214 def remove_notes(lines, notes_group):
igor@39 215 notes_filename = notes_filenames()[0]
igor@39 216 notes = {}
igor@39 217 for k in notes_group.keys():
igor@39 218 if notes_filename in notes_group[k]:
igor@39 219 notes[k] = notes_group[k][notes_filename]
igor@39 220
igor@39 221 result = []
igor@39 222 for line in lines:
igor@39 223 line = line.rstrip('\n')
igor@39 224 match_object = re.match('(\s+)(\S+)(\s+)(\S+)(\s+)(.*)', line)
igor@39 225 if match_object:
igor@39 226 result.append("".join([
igor@39 227 match_object.group(1),
igor@39 228 match_object.group(2),
igor@39 229 match_object.group(3),
igor@39 230 match_object.group(4),
igor@39 231 "\n"
igor@39 232 ]))
igor@39 233 notes[match_object.group(4)] = match_object.group(6)
igor@39 234 else:
igor@39 235 result.append(line+"\n")
igor@39 236
igor@39 237 save_notes(notes_filename, notes)
igor@39 238 return result
igor@39 239
igor@39 240 def save_notes(filename, notes):
igor@39 241 lines = []
igor@39 242 saved_words = []
igor@39 243 with codecs.open(filename, "r", "utf-8") as f:
igor@39 244 for line in f.readlines():
igor@39 245 (word, note) = re.split('\s+', line.rstrip('\n'), maxsplit=1)
igor@39 246 if word in notes:
igor@39 247 line = "%-29s %s\n" % (word, notes[word])
igor@39 248 saved_words.append(word)
igor@39 249 lines.append(line)
igor@39 250 for word in [x for x in notes.keys() if not x in saved_words]:
igor@39 251 line = "%-29s %s\n" % (word, notes[word])
igor@39 252 lines.append(line)
igor@39 253
igor@39 254 with codecs.open(filename, "w", "utf-8") as f:
igor@39 255 for line in lines:
igor@39 256 f.write(line)
igor@39 257
igor@39 258
igor@38 259 def substract_dictionary(dict1, dict2):
igor@38 260 """
igor@38 261 returns dict1 - dict2
igor@38 262 """
igor@38 263 result = {}
igor@38 264 for (k,v) in dict1.items():
igor@38 265 if not k in dict2:
igor@38 266 result[k] = v
igor@38 267 return result
igor@38 268
igor@38 269 def dump_words(words, filename):
igor@38 270 with codecs.open(filename, "w+", "utf-8") as f:
igor@38 271 for word in words.keys():
igor@38 272 f.write(("%s\n"%word)*words[word])
igor@38 273
igor@38 274 def error_message(text):
igor@38 275 print text
igor@38 276
igor@40 277 def find_wordgroups_weights(word_pairs, normalizator):
igor@38 278 weight = {}
igor@40 279 for (num, word) in word_pairs:
igor@38 280 normalized = normalizator.normalize(word)
igor@38 281 weight.setdefault(normalized, 0)
igor@40 282 weight[normalized] += num
igor@38 283 return weight
igor@38 284
igor@38 285 def find_linked_words(notes):
igor@38 286 linked_words = {}
igor@38 287 for word in notes.keys():
igor@38 288 for note in notes[word].values():
igor@38 289 if "@" in note:
igor@38 290 result = re.search(r'\@(\S*)', note)
igor@38 291 if result:
igor@38 292 main_word = result.group(1)
igor@38 293 if main_word:
igor@38 294 linked_words[word] = main_word
igor@38 295 return linked_words
igor@38 296
igor@40 297 def compare_word_pairs(pair1, pair2, wgw, normalizator, linked_words):
igor@40 298 (num1, word1) = pair1
igor@40 299 (num2, word2) = pair2
igor@38 300
igor@38 301 normalized_word1 = normalizator.normalize(word1)
igor@38 302 normalized_word2 = normalizator.normalize(word2)
igor@38 303
igor@38 304 cmp_res = cmp(wgw[normalized_word1], wgw[normalized_word2])
igor@38 305 if cmp_res != 0:
igor@38 306 return cmp_res
igor@38 307 else:
igor@38 308 cmp_res = cmp(normalized_word1, normalized_word2)
igor@38 309 if cmp_res != 0:
igor@38 310 return cmp_res
igor@38 311 else:
igor@38 312 return cmp(int(num1), int(num2))
igor@38 313
igor@40 314 def print_words_sorted(word_pairs, stats, print_stats=True, stats_only=False):
igor@40 315 if stats_only:
igor@40 316 codecs.getwriter("utf-8")(sys.stdout).write("stat_only")
igor@40 317 return
igor@38 318
igor@40 319 if print_stats:
igor@40 320 codecs.getwriter("utf-8")(sys.stdout).write(
igor@40 321 "# %(language)s, %(percentage)s, <%(total_known)s/%(total)s>, <%(groups)s/%(words)s>\n" % stats)
igor@38 322
igor@40 323 level_lines = range(int(float(stats['percentage']))/5*5+5,95,5)+range(90,102)
igor@40 324 known = int(stats['total_known'])
igor@40 325 total = int(stats['total'])
igor@40 326 current_level = 0
igor@40 327 for word_pair in word_pairs:
igor@40 328 codecs.getwriter("utf-8")(sys.stdout).write("%10s %s\n" % word_pair)
igor@40 329 known += word_pair[0]
igor@40 330 if 100.0*known/total >= level_lines[0]:
igor@40 331 current_level = level_lines[0]
igor@40 332 while 100.0*known/total > level_lines[0]:
igor@40 333 current_level = level_lines[0]
igor@40 334 level_lines = level_lines[1:]
igor@40 335 codecs.getwriter("utf-8")(sys.stdout).write("# %s\n" % current_level)
igor@38 336
igor@39 337 def filter_add_notes(args):
igor@39 338 lines = readlines_from_file(args[0])
igor@39 339 notes = load_notes(notes_filenames())
igor@39 340 lines = add_notes(lines, notes)
igor@39 341 with codecs.open(args[0], "w", "utf-8") as f:
igor@39 342 for line in lines:
igor@39 343 f.write(line)
igor@39 344
igor@39 345 def filter_remove_notes(args):
igor@39 346 lines = readlines_from_file(args[0])
igor@39 347 notes = load_notes(notes_filenames())
igor@39 348 lines = remove_notes(lines, notes)
igor@39 349 with codecs.open(args[0], "w", "utf-8") as f:
igor@39 350 for line in lines:
igor@39 351 f.write(line)
igor@39 352
igor@40 353 def filter_get_words_group_words_add_stat(args):
igor@40 354 vocabulary = load_vocabulary()
igor@40 355 notes = load_notes(notes_filenames())
igor@40 356 lines = readlines_from_stdin()
igor@40 357 words = get_words(lines)
igor@40 358
igor@40 359 stats = {}
igor@40 360 stats['total'] = sum(words[x] for x in words.keys())
igor@40 361 words = substract_dictionary(words, vocabulary)
igor@40 362
igor@40 363 stats['total_unknown'] = sum(words[x] for x in words.keys())
igor@40 364 stats['total_known'] = stats['total'] - stats['total_unknown']
igor@40 365 stats['percentage'] = "%7.2f"%(100.0*stats['total_known']/stats['total'])
igor@40 366 stats['groups'] = 0
igor@40 367 stats['words'] = len(words)
igor@40 368 stats['sentences'] = 0 #FIXME
igor@40 369 stats['language'] = config['language']
igor@40 370
igor@40 371 linked_words = find_linked_words(notes)
igor@40 372 normalizator = Normalizator(config['language'], linked_words)
igor@40 373
igor@40 374 word_pairs = []
igor@40 375 for k in sorted(words.keys(), key=lambda k: words[k], reverse=True):
igor@40 376 word_pairs.append((words[k], k))
igor@40 377
igor@40 378 wgw = find_wordgroups_weights(word_pairs, normalizator)
igor@40 379 word_pairs = sorted(
igor@40 380 word_pairs,
igor@40 381 cmp=lambda x,y:compare_word_pairs(x,y, wgw, normalizator, linked_words),
igor@40 382 reverse=True)
igor@40 383
igor@40 384 print_words_sorted(word_pairs, stats)
igor@40 385
igor@37 386 (options, args) = parser.parse_args()
igor@38 387 if options.language:
igor@38 388 config['language'] = options.language
igor@37 389
igor@38 390 if options.function:
igor@38 391 function_names = {
igor@39 392 'add_notes' : filter_add_notes,
igor@39 393 'remove_notes': filter_remove_notes,
igor@40 394 'get_words_group_words_add_stat': filter_get_words_group_words_add_stat,
igor@38 395 }
igor@38 396 if options.function in function_names:
igor@38 397 function_names[options.function](args)
igor@38 398 else:
igor@38 399 error_message("Unkown function %s.\nAvailable functions:\n%s" % (
igor@38 400 options.function, "".join([" "+x for x in sorted(function_names.keys())])))
igor@38 401 sys.exit(1)
igor@37 402
igor@37 403
igor@37 404
igor@37 405
igor@38 406 #os.system("vim")
igor@37 407